Neural Speaker Embeddings for Ultrasound-based Silent Speech Interfaces.
Amin Honarmandi ShandizLászló TóthGábor GosztolyaAlexandra MarkóTamás Gábor CsapóPublished in: CoRR (2021)
Keyphrases
- speech recognition
- speaker recognition
- audio visual
- automatic speech recognition
- speaker verification
- speaker identification
- speech signal
- prosodic features
- speaker dependent
- vocal tract
- speaker diarization
- network architecture
- automatic speech recognition systems
- ultrasound images
- speech synthesis
- spoken dialogue systems
- neural network
- multimodal interfaces
- speaker adaptation
- noisy environments
- multi modal
- speech sounds
- synthesized speech
- acoustic features
- speech recognizer
- emotion recognition
- hands free
- broadcast news
- dimensionality reduction
- probabilistic neural network
- user interface
- gaussian mixture model
- neural model
- audio stream
- low dimensional
- phoneme recognition
- hidden markov models
- euclidean space
- vector space
- visual information
- automatic transcription
- manifold learning
- acoustic models
- text to speech
- language model
- associative memory
- pattern recognition
- distance measure
- speech recognition systems
- recognition engine
- spoken document retrieval
- visual data
- interface design